A self regulating insulin delivery system has been designed based on the concept of competitive binding between a glycosylated insulin derivative (SAPG-insulin) and glucose to a Concanavalin A (Con A) polymeric gel substrate. The competitive binding of the two ligands for the substrates will regulate the glycosylated insulin release, in relation to the outside glucose concentration, while the polymeric membranes, serving as a pouch system containing glycosylated insulin and Con A gel, are used to regulate the permeability of glucose influx and glycosylated insulin efflux. Succinylamido-phenyl-glucopyranoside coupled insulin (SAPG- insulin) will continue to be used, based on its simple synthetic procedure and the well characterized data. In addition, SAPG-insulin was found to be non-immunogenic, bioactive and pharmacodynamically active; i.e., comparable to commercial insulin. Previous in vitro and in vivo experiments have demonstrated the success of this device and the need for further optimization research. The Con A protein molecule will be crosslinked to obtain a bioactive (preserved binding sites) macromolecule. This design will inhibit the leakage of the substrate from the pouch, in order to prevent immunological reactions. An adequate polymeric membrane, which has similar permeability of glucose influx and SAPG-insulin efflux, will be utilized for the design of this pouch system suing a heat sealing process. The fabricated pouch, containing SAPG-insulin and Con A gel, will be studied in vitro for SAPG-insulin release versus outside glucose concentration and for the time it takes the device to respond to the change in glucose concentrations (lag time). The membrane itself and as the complete pouch system will be tested in terms of in vivo peritoneal biocompatibility (fibrous growth, foreign body invasion, etc.). The final system, following optimization (1-3 month period), will be implanted in pancreatectomized dog models. The system will be recharged using a subcutaneous needle every three months during an experiment period of 9 months to one year. Diabetic dogs with this device will be carefully monitored in terms of metabolic balance and biochemical evaluation. The obtained data should provide criteria for successful human application after completion of the proposed four year period.